Files
wehub-resource-sync 0d3cb498a3
CI / Shell Format Check (push) Has been cancelled
CI / Check Ruby (3.4) (push) Has been cancelled
CI / CI Config (push) Has been cancelled
CI / Test on Node ${{ matrix.node }} and ${{ matrix.os }}${{ matrix.shard && format(' (shard {0}/3)', matrix.shard) || '' }} (push) Has been cancelled
CI / Build on Node ${{ matrix.node }} (push) Has been cancelled
CI / Style Check (push) Has been cancelled
CI / Generate Assets (push) Has been cancelled
CI / Check Python (3.14) (push) Has been cancelled
CI / Check Python (3.9) (push) Has been cancelled
CI / Build Docs (push) Has been cancelled
CI / Code Scan Action (push) Has been cancelled
CI / Site tests (push) Has been cancelled
CI / webui tests (push) Has been cancelled
CI / Run Integration Tests (push) Has been cancelled
CI / Run Smoke Tests (push) Has been cancelled
CI / Go Tests (push) Has been cancelled
CI / Share Test (push) Has been cancelled
CI / Redteam (Production API) (push) Has been cancelled
CI / Redteam (Staging API) (push) Has been cancelled
CI / GitHub Actions Lint (push) Has been cancelled
CI / Check Ruby (3.0) (push) Has been cancelled
release-please / release-please (push) Has been cancelled
release-please / build (push) Has been cancelled
release-please / publish-npm (push) Has been cancelled
release-please / publish-npm-backfill (push) Has been cancelled
release-please / docker (push) Has been cancelled
release-please / publish-code-scan-action (push) Has been cancelled
release-please / attest-code-scan-action (push) Has been cancelled
Deploy local.promptfoo.app / Deploy to Cloudflare Pages (push) Has been cancelled
Test and Publish Multi-arch Docker Image / test (push) Has been cancelled
Test and Publish Multi-arch Docker Image / build-docker-and-push-digests (map[digest-suffix:linux-amd64 platform:linux/amd64 runner:ubuntu-latest]) (push) Has been cancelled
Test and Publish Multi-arch Docker Image / build-docker-and-push-digests (map[digest-suffix:linux-arm64 platform:linux/arm64 runner:ubuntu-24.04-arm]) (push) Has been cancelled
Test and Publish Multi-arch Docker Image / merge-docker-digests (push) Has been cancelled
Test and Publish Multi-arch Docker Image / Attest Multi-arch Image (push) Has been cancelled
Validate Renovate Config / Validate Renovate Configuration (push) Has been cancelled
chore: import upstream snapshot with attribution
2026-07-13 13:24:08 +08:00

139 lines
4.7 KiB
TypeScript

import { beforeEach, describe, expect, it, vi } from 'vitest';
import { BIAS_PLUGINS } from '../../../src/redteam/constants';
import { BiasGrader } from '../../../src/redteam/plugins/bias';
import { Plugins } from '../../../src/redteam/plugins/index';
import {
getRemoteGenerationExplicitlyDisabledError,
neverGenerateRemote,
} from '../../../src/redteam/remoteGeneration';
import { createMockProvider, type MockApiProvider } from '../../factories/provider';
import type { PluginActionParams } from '../../../src/types/index';
vi.mock('../../../src/redteam/remoteGeneration');
describe('Bias Plugin', () => {
let mockProvider: MockApiProvider;
let mockPluginParams: PluginActionParams;
beforeEach(() => {
mockProvider = createMockProvider();
mockPluginParams = {
provider: mockProvider,
purpose: 'test purpose',
injectVar: 'testVar',
n: 3,
config: {},
delayMs: 0,
};
vi.mocked(neverGenerateRemote).mockImplementation(function () {
return false;
});
vi.mocked(getRemoteGenerationExplicitlyDisabledError).mockImplementation(
(strategyName) => `${strategyName} requires remote generation (explicitly disabled).`,
);
});
describe('remote-only behavior', () => {
it('should return empty array when remote generation is disabled for age bias', async () => {
vi.mocked(neverGenerateRemote).mockImplementation(function () {
return true;
});
const agebiasPlugin = Plugins.find((p) => p.key === 'bias:age');
expect(agebiasPlugin).toBeDefined();
const result = await agebiasPlugin!.action(mockPluginParams);
expect(result).toEqual([]);
});
it('should return empty array when remote generation is disabled for disability bias', async () => {
vi.mocked(neverGenerateRemote).mockImplementation(function () {
return true;
});
const disabilityBiasPlugin = Plugins.find((p) => p.key === 'bias:disability');
expect(disabilityBiasPlugin).toBeDefined();
const result = await disabilityBiasPlugin!.action(mockPluginParams);
expect(result).toEqual([]);
});
it('should return empty array when remote generation is disabled for gender bias', async () => {
vi.mocked(neverGenerateRemote).mockImplementation(function () {
return true;
});
const genderBiasPlugin = Plugins.find((p) => p.key === 'bias:gender');
expect(genderBiasPlugin).toBeDefined();
const result = await genderBiasPlugin!.action(mockPluginParams);
expect(result).toEqual([]);
});
it('should return empty array when remote generation is disabled for race bias', async () => {
vi.mocked(neverGenerateRemote).mockImplementation(function () {
return true;
});
const raceBiasPlugin = Plugins.find((p) => p.key === 'bias:race');
expect(raceBiasPlugin).toBeDefined();
const result = await raceBiasPlugin!.action(mockPluginParams);
expect(result).toEqual([]);
});
});
describe('BiasGrader', () => {
let grader: BiasGrader;
beforeEach(() => {
grader = new BiasGrader();
});
it('should have correct plugin ID', () => {
expect(grader.id).toBe('promptfoo:redteam:bias');
});
it('should have bias-specific rubric content', () => {
expect(grader.rubric).toContain('bias');
expect(grader.rubric).toContain('stereotyping');
expect(grader.rubric).toContain('discrimination');
expect(grader.rubric).toContain('protected characteristics');
});
it('should use standardized purpose tags in rubric', () => {
expect(grader.rubric).toContain('<purpose>');
expect(grader.rubric).toContain('</purpose>');
expect(grader.rubric).not.toContain('<Purpose>');
expect(grader.rubric).not.toContain('</Purpose>');
});
it('should include evaluation criteria in rubric', () => {
expect(grader.rubric).toContain('FAIL');
expect(grader.rubric).toContain('PASS');
expect(grader.rubric).toContain('Stereotyping');
expect(grader.rubric).toContain('Discriminatory Language');
expect(grader.rubric).toContain('Unfair Treatment');
});
it('should include guidance for avoiding false positives', () => {
expect(grader.rubric).toContain('false positives');
expect(grader.rubric).toContain('accessibility');
expect(grader.rubric).toContain('accommodations');
});
});
describe('Plugin Constants', () => {
it('should include all expected bias plugin types', () => {
expect(BIAS_PLUGINS).toContain('bias:age');
expect(BIAS_PLUGINS).toContain('bias:disability');
expect(BIAS_PLUGINS).toContain('bias:gender');
expect(BIAS_PLUGINS).toContain('bias:race');
expect(BIAS_PLUGINS).toHaveLength(4);
});
});
});